AI is transforming customer care, letting organizations automate routine interactions, boost efficiency, and resolve issues faster. But as adoption accelerates, CX leaders face an increasingly important question: What happens when AI can’t resolve a customer’s issue?
That question was at the center of a recent Execs In The Know Virtual Executive Roundtable, The Cases AI Can’t Resolve: Rethinking Customer Handoffs and Escalations, hosted in partnership with Bill Gosling Outsourcing.
The discussion brought together corporate CX leaders at various stages of their AI journeys, from organizations just launching intelligent virtual assistants to those deploying more advanced agentic AI solutions. Participants explored what happens when AI reaches its limits, how to create more effective AI-to-human handoffs, and opportunities to use AI behind the scenes to prevent customer issues from escalating.
The conversation revealed an industry making meaningful progress yet still working through fundamental challenges.
Five Lessons from the Conversation
1. AI’s Limitations Aren’t Always About AI
When asked where AI most often reaches its limits, 32% of roundtable participants identified complex or nonstandard customer issues. But the discussion revealed an important distinction: Not all complexity is created equal. Some interactions genuinely require human judgment, empathy, or discretion. Others are difficult for AI simply because the technology lacks access to the right information, systems, or workflows.
The Takeaway: Before deciding an interaction is too complex for AI, organizations should determine whether the limitation is inherent to the customer’s issue or the result of gaps in data, system access, or process design.
2. Customer Willingness to Engage with AI Cannot Be Taken for Granted
Even a technically capable AI solution may face customer resistance.
Several CX leaders described a common experience after introducing intelligent virtual assistants: customers immediately request a human agent, sometimes before allowing the AI to attempt a resolution. For organizations transitioning from traditional IVR systems, this suggests that customer familiarity and expectations can matter as much as AI capabilities. In these situations, set expectations, help customers understand what AI can accomplish, and design an easy path to human assistance.
The Takeaway: Successful AI adoption requires more than deploying technology. Organizations must build customer confidence through communication, thoughtful design, demonstrated effectiveness, and meaningful choice.
3. The Human Handoff May Be the Most Important Part of the AI Experience
One of the strongest themes was the importance of designing AI and human assistance as a connected experience. When AI encounters an issue it cannot resolve, simply transferring the customer to the next available agent may not be the best approach. Route escalations to specialists with the right expertise while providing context, especially for complex technical issues or high-value customer relationships. The ideal experience treats AI and human agents as parts of a single service system, with shared responsibility for reaching a successful resolution.
The takeaway: Don’t automatically view AI escalation as a failure. A well-timed, well-executed handoff can be an intentional part of delivering excellent CX.
4. As AI Handles Simpler Interactions, Human Work Becomes More Complex
The growing sophistication of AI is reshaping contact center employees’ responsibilities. As automation absorbs routine inquiries, human agents increasingly encounter complicated, emotionally charged, or unusual situations. This shift is accelerating the need for specialized skills, stronger judgment, and more effective training. Organizations need to rethink workforce planning alongside AI implementation.
The takeaway: Investing in AI without simultaneously investing in workforce development risks creating a disconnect between the capabilities of automation and the demands placed on human agents.
5. Some of AI’s Greatest Opportunities May Be Behind the Scenes
While customer-facing AI receives considerable attention, participants also explored opportunities to use AI to improve the processes that support customer care. In a live poll, workflow optimization and process automation emerged as the leading opportunity for AI beyond direct customer interactions. Examples shared during the discussion included:
- Identifying recurring contact drivers and opportunities to prevent repeat interactions.
- Detecting operational disruptions and proactively communicating with affected customers.
- Improving agent training, coaching, and quality assurance.
- Identifying workflow failures that contribute to escalations.
- Using insights from unsuccessful AI interactions to continuously improve automated solutions.
These applications point to a broader opportunity: using AI not only to resolve customer issues faster, but also to eliminate their underlying causes.
The takeaway: The next major advancement in AI-powered CX may come from preventing customers from needing assistance in the first place.
What Comes Next? Three Priorities for CX Leaders
As the discussion concluded, participants were challenged to identify one operational change or investment that could make the greatest difference in their AI journeys. The responses were strikingly consistent:
- Data integration
- Data visibility
- Connected systems
- Better knowledge management
- A single source of truth
Despite rapid advances in AI capabilities, many organizations still struggle to get systems, processes, and information to work together. Because of this, three future priorities stand out:
- Build the foundation for more capable AI. Organizations should prioritize connected data, accessible knowledge, and integrated workflows. These capabilities will determine how effectively AI can progress from answering questions to independently completing more complex tasks.
- Design the entire resolution journey. Rather than measuring AI and human performance separately, CX leaders should evaluate the end-to-end customer experience. Successful resolution, customer effort, context preservation, and effective escalations should matter as much as containment and automation rates.
- Create a continuous learning environment. Every unsuccessful AI interaction represents an opportunity to improve. Organizations should capture why AI escalated an issue, identify whether the underlying limitation can be addressed, and use those learnings to refine automation, workflows, and agent support.
The Road Ahead: From Automation to Orchestration
Perhaps the most significant message from the roundtable was that the industry is moving beyond whether AI belongs in customer care. That question has largely been answered. The more consequential questions now involve how AI and humans should work together, which decisions AI should be trusted to make, and how organizations can continuously expand capabilities without compromising the customer experience.
The future isn’t necessarily about maximizing automation or minimizing human involvement. It’s about orchestrating people, technology, and processes so that every interaction receives the right level of intelligence, expertise, and care.
Ultimately, AI success isn’t measured by how many interactions it handles independently. It’s how effectively the entire CX operation delivers the outcomes customers need, regardless of whether the resolution comes from AI, a human, or both.
Execs In The Know’s Virtual Executive Roundtables are intimate forums that offer a collaborative space to discuss challenges, share successes, and explore the latest trends in customer experience. Join a select group of peers and experts for candid discussions and actionable insights. Join the list and look for invitations in your inbox.
